Robotic devices have been used in surgery since the late 1980s, beginning with a serial kinematic structured industry robot programmed to position a tool guide at a specified location near the head [Kwoh 1988] [Lavalée 1989]. During the 1990s, robots were introduced into orthopedic surgery with the ROBODOC system (ISS, USA) and the CASPAR (U.R.S., Germany) for hip and knee surgery [Bargar 1998, Prymka 2006]. However, not only did these autonomous systems not show more long term advantages than conventional techniques, they also showed longer operation times and increased blood loss [Bach 2002]. One of the disadvantages of these systems was the rigid fixation of the bone to the robot and the fully autonomous processing which took control away from the surgeon. Autonomous robots are not well-suited for soft-tissue surgery, as the shape of tissue may change when cut or pushed, or as a result of the patient's breathing.
Master-slave controlled robots, controlled via a console and remote visual feedback, have also been introduced [Maeso 2010]. The da Vinci system is used in minimally-invasive laparoscopic abdominal surgery, where the surgeon controls up to four robotic arms and the movements of the surgeon's hand can be filtered and scaled to enable precise instrument micro-movements.
In contrast to the da Vinci approach, where the surgeon controls the robot from a distance using a remote console, robots have also been introduced for cooperative work. These systems either position a tool guide and the surgeon itself guides the instrument [Liebermann 2006, Plaskos 2005], or the surgeon guides a haptic-controlled robot, and the robot prevents access to forbidden areas.
The concept of the haptic-controlled robot using “active constraints” or “virtual fixtures” was first implemented by the Acrobot system [Davies 2007, Yen 2010], and later by the MAKO Surgical Corp. RIO system. for UKA and THA [US 2006/0142657 A1 (Quaid et al.), Lonner 2010, Dorr 2011]. A randomized prospective study of the Acrobot system showed that with robotic bone preparation in UKA, the tibiofemoral alignment was within 2° of the planned position, whereas in the control group only 40% were below 2° [Cobb 2006].
A similar approach, but without using a huge robot system, is the use of “intelligent” high speed burrs that can be programmatically enabled and disabled in pre-planned areas. Whereas the Navigated Control concept controls the rotating speed of the burr [Strauss 2005], the Precision Freehand Sculptor comprises a burr that retracts behind a guard [Brisson 2004, WO 2011/133927 A2 (Nikou et al.)].
US 2005/0171553 (Schwarz et al.) discloses a handheld device for treating a body part that comprises a base, a tool that is able to move with respect to the base, and an actuation unit to move the tool within a predetermined working space to a predetermined position on the part to be treated. This device takes into account the movements of the base and of the part to be treated by detecting the position of the tool and the position of the part to be treated, by comparing said positions with a target position and by adapting the actuation unit accordingly. US 2012/0143084 (Shoham) describes a handheld device for treating a body part that comprises a handle, a tool that is able to move with respect to the handle, and a robot to move the tool within a predetermined working space to a predetermined position on the part to be treated. This device is able to detect the position of the tool with respect to a forbidden zone and to change the pose of the robot if the user moves the handle by an amount that would bring the tool within said forbidden zone. A disadvantage of these hand-guided tools is that the milling path itself needs to be controlled by the surgeon, which may result in non-efficient bone removal, inaccurate milling surfaces and unintended heat emission. In addition, significant time is required to obtain reasonable overall accuracy. Locally, milling surfaces are always bumpy or irregular, and to compensate for such bumps, cement is usually required between the milling surface and the implant, instead of non-cemented implants which are often preferred. Even with cement, the end result of such a process is an overall loss of accuracy.
For applications where no tremor is permitted, such as retinal surgery, a steady hand manipulator has been introduced [Mitchell 2007, Uneri 2010] as well as the handheld actively stabilized Micron device [Becker 2011, MacLachlan 2012].
US 2011/0208196 (Radermacher et al.) discloses a handheld reactive device for creating and applying constraints to the user that comprises a handle, a tool that is able to move with respect to the handle, and a support element connected to the handle by means of which the handle may be supported on a body surface. The support element is movable via an actuation unit which enables tool repositioning with respect to body surface by shifting the support element, based on sensor data obtained during the treatment. However, this device is applying constraints to the user depending on the material being worked, and it is not able to optimize the tool path actively in view of treating a planned volume of the body.
There remains a need for a lightweight, handheld surgical device which enables tool path optimization and compensation of the surgeon's small movements to minimize vibration and optimize accuracy.